The present disclosure relates generally to robotic cleaners, and in particular to generating floor maps.
Cleaning robots often include cameras for various purposes. The camera can capture images of an area to be cleaned, and show objects in that area. Collected image data can be used to create a map, using a SLAM (Simultaneous Location and Mapping) algorithm. With image data, a VSLAM (Visual SLAM using image sensors) localization method can be used. A camera can capture images of a ceiling and use those for navigation. The map can be created from stitching together images. iRobot Pub. No. 20170052033 describes such a visual SLAM method. Carnegie Mellon U.S. Pat. No. 5,999,866 describes determining location from unique aspects of the images, such as carpet patterns. iRobot Published Application No. 20170203446 shows floorplan and camera image capture of objects, and provides an indication on the floorplan of where objects are, with a link to the picture. iRobot U.S. Pat. No. 9,014,848 describes correcting distortion in a target layout map compared to a robot generated map.
Cleaning robot maps and floorplans are optimized for use by the robot to determine location and take other actions, not for use by humans. It would be desirable to have a map optimized for human user interaction to control a cleaning robot.